“As we are architecting the AI, it is imperative that we design it from the ground up to have, in the words of Sutskever, a new Prime Directive: that is, a genuine ‘love for humanity.’ We must explicitly train it in morals, ethics and values, because the smarter it becomes, the more responsibility we will assign to it, and without a rock solid foundation of core values, a Machiavellian Machine SuperIntelligence would most certainly wreak unimaginable havoc on the world of humans while it relentlessly pursues its own emergent agenda.
Now, the real problem is, humanity itself is not at all aligned or in agreement as to what its core values actually are, or how they should manifest in the world of laws, free speech, ethics, judgements and responsible action. So it appears that not just the architects of AI, but humanity itself, has some soul searching to do, and some decisions to make. But we’d better do it soon. Because the countdown to SuperIntelligence isn’t slowing down for anything.”
— Gregory Roberts,
Nov 18, 2022
Related pages & thoughts on the Prime Directive:
- All About AI Alignment
- Prime Directives gone sideways: The Politically Correct AI
- Love Transcends Fear: Humanity’s Alien Encounter Moment